...
\uD83D\uDDE3 Discussion topics
Item | Presenter | Notes |
---|---|---|
Opening | Jeremy | Jeremy welcomed the participants and presented the agenda:
|
1 Review WIS2 Pilot Phase Plan, check progress | Hassan Enrico | Hassan updated on the progress of wis2 pilot |
phase:
For Cg-19, |
Task Team on NWPMD
Q1:Whether data other than core data can be cached by GC? (mandatory GDPFS products are not defined as core data now)
(Baudouin) Caching is different from Data Policy.
(Remy) Q3: Mechanismwe plan to have a booth and maybe a side event | ||
Task Team on NWPMD questions | Hassan | Hassan reported on some questions raised by TT-NWPMD:
(Baudouin) don’t mix the caching with the Data Policy, technically they can be different. (Remy) recommended data at the moment shouldn't be cached. and asked if are we going to cache all core data? that was already discussed when Baudouin mentioned the big data from ECMWF. Remy refer to the issue on GitHub - Prevent caching of large datasets in Global Cache (https://github.com/wmo-im/wis2-guide/issues/7 |
Q2: Can NWP centers bundle core data and other data?
Anna> The definition of core is still very much under discussion. Like what exactly? in the data policy, there are very broad statements about what the types of data are, and that could fit many different types of things under these broad definitions. So I feel like that shouldn't limit somebody from sharing the data on the global cache if if they're still working on the definitions of what core is. Jeremy> if that recommended data is is huge, then as a global cache operator I'm probably going to say actually I can't afford to to cache your data for you so.
Baudouin> it’s not limited to NWP, it’s true for any data (Remy) |
at the moment we don’t have any mechanism. Shall we have one or just publish the new dataset, with a new message and new ID? Jeremy>In our pilot system, we need to have a use case which is testing what happens when people want to replace a bad file and see how client applications respond. (Weiqing) in GTS, header remains the same, message with correction (COR). (Baudouin) Agrees that the current GTS way |
of dealing with the updated data (with a correction flag) is good. (Remy): Proposed to create an issue of the notification message on the GitHub Action: Hassan to create an issue on that repository explaining the use case or explaining the requirement on message notification mechanism on updated data. |
Q3: WIS to clarify if a part of a big file (of specific variable(s) or of some sub-domain(s)) can be extracted using API prepared by Global |
Service. (Jeremy) Currently, no interactive API available. GC only takes the files, dump the data for the user to download. ( |
Baudouin) URL is one kind of API. Serving data via API by Global Services was not fully discussed. |
(Remy) Emphasizes that API service (added value service) will |
not be provided by the Global Services. It is the Members or external parties to provide additional added value services. Q5: WIS to clarify how subscription to the broker will work; particularly, the possibility on filtering certain kinds of products under a particular topic |
(Remy)
(Anna)
(Baudouin) In NWP, e.g., for each leap time, there is one file, including all the parameters.
(Jeremy) discovery metadata(e.g. the possibility of only getting notifications on tropical cyclone tracks). (Remy) the answer is simple. If it's in the topic hierarchy, you can filter. So if it makes sense for the NWP colleagues to define the topic level saying tropical cyclone tracks as part of the topic hierarchy featuring, it's easy and if not, it can be done. | ||
2. wis2box update | Enrico | We are working on a number of issues. After training, we discover some issues. Synoptic is working on UI design. |
3. Prioritise and review open GitHub issues | Jeremy | https://github.com/wmo-im/wis2pilot/issues open issues but without update. (Jeremy) Propose to add wis2 pilot phase on the WIS community platform activity area. (Remy) WIS to GTS Gateway & GTS to WIS Gateway is still not clear. Indonesia request to play GB replacing Australia is not updated. (Hassan) To collect the questions and progress of each issue from the participants, either by updating GitHub repository (Jeremy) or a compiled PDF (Remy). (Remy) end of June, Cache is available from USA. (Weiqing) We are still working on Global Cache. For Global Broker, our decision remains the same. We could also do Global Broker if it is insisted. Decision: It would be better that Australia works as a Global Broker instead of Global Cache considering there are four players for GC. Action: Hassan to help coordinate about the general GC issues. Update on each WP1-11 Action: Hassan to compile the progress on wis2nodes. (Timo)
(Hassan) We plan to establish a task team (expert team?) (most of them are from ICAO) and we propose to choose from TT-AvData as the candidates as some of new TT members are from Met Office. |
✅ Action items
- HADDOUCH Hassan to create an issue on wis2pilot repository about the mechanism of message notification of new version of data file.
- HADDOUCH Hassan Anna Milan to track the issues from TT-NWPMD on the wis2pilot repository
- HADDOUCH Hassan to present a progress report on all the wis2node at the next meeting
- HADDOUCH Hassan to collect the questions and progress of each issue from the participants in WIS2 pilot phase, either by updating GitHub repository (Jeremy’s preference) or a compiled PDF (work for Remy).
...